Cryptococcal Meningitis: The Trial That Says Wait
For most opportunistic infections, earlier antiretroviral therapy saves lives. Cryptococcal meningitis is the one place a randomized trial found the opposite, and the team has to decide how literally to read it.
E.O., a 41-year-old woman, has cut hair at the same strip-mall salon for close to fifteen years and built a loyal client base doing it — regulars who, over the past month, had started asking whether she was feeling all right, since the headaches she'd been blaming on the salon's fluorescent lighting had visibly worsened and she'd twice had to sit down mid-appointment. She came in three days ago after a morning of vomiting and a headache bad enough that light through the blinds made her cry, and was found confused and photophobic in the emergency department. A CT scan showed no mass lesion, and a lumbar puncture returned an opening pressure of 34 cmH2O with cryptococcal antigen strongly positive in both serum and CSF. An HIV test drawn at the same visit came back positive — her first, and by her account her only prior positive test of any kind was a normal mammogram two years ago.
Her CD4 count is 19 cells/µL. She started induction therapy with liposomal amphotericin B and flucytosine yesterday, and a second lumbar puncture this morning brought her opening pressure down to 22 — real but incomplete relief. The question the team is weighing is not whether to start ART, but when. COAT, the trial that actually answers this, randomized patients with HIV-associated cryptococcal meningitis to ART at 1-2 weeks versus 5 weeks after diagnosis and stopped the early-ART arm for a mortality difference — 45% dead by 26 weeks against 30% in the deferred arm, a result driven by IRIS in a compartment, the central nervous system, where inflammatory swelling has nowhere safe to go. Every other OI in her chart's differential — TB, PCP — argues the opposite direction. This one doesn't. Her fungal burden itself is severe by the same measure the team is following serially: a quantitative CSF culture at diagnosis grew greater than 100,000 colony-forming units per milliliter, a number the induction regimen is expected to drive down steadily over the next two weeks, and the pace of that decline, not just the calendar, is part of what the team means when it talks about her CSF clearing.
Neuro-ICU, hour one after the diagnostic lumbar puncture
COAT is the trial that answers this question directly, and it answers it the opposite way from almost every other opportunistic infection. Early ART at 1-2 weeks versus deferred ART at 5 weeks, randomized, mortality endpoint, stopped for harm in the early arm — 45% dead by 26 weeks versus 30%. The mechanism is IRIS in a compartment that doesn't tolerate swelling. Defer ART to 5 weeks.
I'm not disputing COAT's direction — deferred is right, and 1-2 weeks is wrong. I'm questioning whether 5 weeks is itself load-bearing or just the number the trial happened to test.
The trial compared two fixed arms; it didn't establish that 5 weeks is a floor she can't cross earlier if her CSF clears faster than typical. WHO's own updated guidance describes 4-6 weeks as the acceptable window once induction therapy is established and pressure is controlled, not a single mandated day. If her repeat CSF cultures sterilize early and her pressure stays controlled, I'd rather use her own trajectory than a calendar date picked for trial design reasons. And her mental status is already trending the right way on induction therapy alone, before ART has entered the picture at all — that's independent evidence the inflammatory process is moving in the right direction on its own timeline, not a reason to accelerate ART to help it along.
Whatever week ART ends up starting, it isn't the thing that kills her this week. An opening pressure of 34, even down to 22, is still elevated intracranial hypertension, and that alone is independently associated with death and vision loss in cryptococcal meningitis regardless of ART timing. Serial therapeutic lumbar punctures, targeting a closing pressure under roughly half the opening value or under 20, need to continue daily until it stabilizes — that's the nearer-term mortality risk in this room right now.
Agreed without real dispute on direction: ART deferred, targeting week 5, with the exact day left open to her CSF-clearance trajectory rather than fixed today. Serial therapeutic lumbar punctures continue daily until closing pressure stabilizes under roughly half of today's opening value — the pressure-management plan the critical care physician raised, which proceeds identically no matter which week ART eventually starts.